Information on the web is increasing at an enormous speed. Search histories of the users, browsing pattern, query expansion using relevance feedback are some of the techniques used in the literature to personalize the web search. This study proposed the integrated approach of personalized web search using Agents and Information Scent. Agent based information retrieval system personalizes the web search by clustering the query sessions of users on the web using information scent, information scent is the measure of the sense of value of clicked web page in the query session with respect to the information need of the user. Interface agent after receiving the input query generates the query recommendations using the cluster which is closes to the information need of input query and is expanded using related keywords to disambiguate its context. The proposed framework effectively personalizes the web search through input query sense disambiguation and exposes the part of Surface Web through Hubs and Authorities recommendations using user profile modeled using Information Scent of clicked URLs and clusters of query sessions on the web. The effectiveness of the proposed system on the precision of search results is confirmed with the experiments conducted on the data set collected using proposed architecture.